Title:Reactive Trajectory Generation for Multiple Vehicles in Unknown Environments with Wind Disturbances

Abstract:Unmanned aerial vehicle (UAV) use continues to increase, including operating beyond line of sight in unknown environments where the vehicle must autonomously generate a trajectory to safely navigate. In this article, we develop a trajectory generation algorithm for vehicles with second-order dynamics in unknown environments with bounded wind disturbances where the vehicle only relies on its on-board distance sensors and communication with other vehicles to navigate. The proposed algorithm generates smooth trajectories and can be used with high-level planners and low-level motion controllers. The algorithm computes a maximum safe cruise velocity for the vehicle in the environment and guarantees that the trajectory does not violate the vehicle's thrust limitation, sensor constraints, or user-defined clearance radius around other vehicles and obstacles. Additionally, the trajectories are guaranteed to reach a stationary goal position in finite time given a finite number of bounded obstacles. Simulation results demonstrate the algorithm properties through two scenarios: (1) a quadrotor navigating through a moving obstacle field to a goal position, and (2) multiple quadrotors navigating into a building to different goal positions.